A Claude visibility tracker, searches included
Claude searches the web when a question calls for it, and cites what it reads. GEOFast asks Claude your buyers' questions each week through Anthropic's API with web search on, and counts how often it names you, puts you first and cites your site, with the searches it ran.
Weekly on every plan, of a team's first 5 to 25 prompts.
Claude, through Anthropic's API with web search.
Claude Haiku 5.5, with web search
Claude is asked through Anthropic's API, on Claude Haiku 5.5, with Anthropic's web search tool switched on. Our data provider doesn't capture Claude's app, so reports say Claude was read through its API.
At most two searches an answer
GEOFast lets Claude run up to two web searches for each answer. Claude decides whether to search at all, and names each search it runs, so its answers carry their searches.
Searched from the United States
Claude's searches are placed in the United States for every project, whatever country it tracks, so one week's answers compare with the next.
On your first prompts
Claude is asked of a team's oldest active prompts, up to the plan's number for Claude, so the questions you add first are the ones Claude answers.
What Claude did with our questions.
We asked Claude, with web search on, the demo projects' questions, and counted when it searched, what it cited and how steady it was about Plausible.
answers ran a web search, of those whose count Anthropic's API reported; 1.5 searches each on average.
sources per Claude answer, on average. 27 answers cited none.
first-pick rate for Plausible, 21 of 66, likely 23 to 42%, against 77% named at all.
of the sites Claude cited for a question (1,385 of 1,921) were cited by none of the five other engines for it in that run.
Claude named Plausible in 68% to 82% of a run's answers. The ranges overlap, so the dashboard would call none of those moves a change.
191 answers from Claude through Anthropic's API with web search, from GEOFast's two demo projects (four runs for trygeo.fast and three runs for plausible.io, October 6 and 7, 2026), on questions about AI visibility tools and about web analytics, in US English. Plausible's rates leave out the questions that name it.
Limits: the API's answers can differ from what Claude's apps show. The questions came from GEOFast's onboarding, so they aren't measured demand. A site counts once per answer.
Claude on your dashboard.
Weekly rates, with ranges
How often Claude names you, offers you first and cites you, each with its 90% range, from its weekly answers.
The searches Claude ran
Each Claude answer lists the searches it named, the words it chose for your buyers' question. Searches across a project are on Growth, Agency and the trial. Try the free query fan-out tool
What it cited
Claude's citations join the Sources page, with the pages that name a rival and not you marked, on Growth, Agency and the trial.
How many prompts Claude gets.
Watch, $29 a month
Claude gets the first 5 of the plan's 50 prompts, once a week.
Growth, $99 a month
Claude gets the first 10 of the plan's 100 prompts, once a week.
Agency, $249 a month
Claude gets the first 25 of the plan's 1,500 prompts, once a week.
The 14-day trial
Claude gets the first 5 of the plan's 25 prompts, once a week.
What Anthropic says about Claude's search.
Claude chooses when to search
Anthropic says Claude searches when a request depends on current or changing information, such as specific products or organizations, and answers stable facts without searching.[1]
Citations are always on
With web search, every answer comes with citations to the pages it drew from, and a search can be localized with an approximate user location.[1]
Claude-SearchBot indexes for search
Anthropic says blocking Claude-SearchBot stops it indexing your content for search, which may reduce your visibility in Claude's search results.[2] Claude-SearchBot
Claude-User fetches for a person
Blocking Claude-User stops Claude retrieving your pages in response to a question, which Anthropic says may reduce your visibility in user-directed search. ClaudeBot is the one for training.[2] Claude-User · ClaudeBot
Questions about tracking Claude.
Why does GEOFast ask Claude only some of my prompts?
Claude answers only through its API, and each answer costs several times a captured one. So every plan asks Claude weekly of a set number of a team's first prompts, which keeps the plans' prices.
Which of my prompts does Claude answer?
The team's oldest active ones, up to the plan's number for Claude. Pause or remove one of them and the next prompt in line takes its place.
Is Claude's app the same as what GEOFast asks?
Not quite. GEOFast asks Anthropic's API on one model with web search on, the same way each week. The apps can run other models and settings, so their answers can differ.
Does Claude always search before answering?
No. Claude decides from the question, and Anthropic says it answers stable facts without searching. Buyers' questions about products usually do get a search, and those searches show on the answer.
Where each fact comes from.
Each numbered source is the page a fact comes from, most of them the documentation of the company behind Claude, read on October 21, 2026. These pages change, so check them before you act.
- 1Anthropic: Web search toolplatform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool
- 2Anthropic: Does Anthropic crawl data from the web, and how can site owners block the crawler?support.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler
Keep reading
- ChatGPT rank tracker, counted as first picksTrack your rank in ChatGPT as a first-pick rate with a 90% range, read from the ChatGPT app signed out in your country, with the searches it ran. From $29.
- Perplexity rank tracking, weekly, with rangesTrack how often Perplexity names you, puts you first and cites your site, asked weekly through its Sonar API with search, with a 90% range on each rate.
- Gemini visibility tracker, from the Gemini appTrack how often Google's Gemini app names your brand, puts it first and links your site, captured signed out in your country, with a 90% range on each rate.
- Google AI Mode tracking, apart from OverviewsTrack how often Google's AI Mode names your brand and links your site, read from Google's results in your country and counted apart from AI Overviews.
- AI visibility trackingTrack how often ChatGPT, AI Overviews, AI Mode, Gemini, Perplexity and Claude name you, put you first and cite your site, with a 90% range on every rate.
- The free Query fan-out tool for ChatGPT and ClaudeType a question and see the web searches ChatGPT and Claude ran to answer it, with the pages each cited. Free, five questions a day.
- PricingThe plans, what each includes, and the trial.
Find out what AI says about your site.
Paste your URL and leave with a ranked list of what to fix, with steps and code.
- Free
- No account